gpt-oss-safeguard-20b
OPENAI Developer Architecture Profile
Intelligence (ELO)1065Chatbot Arena Verified
Max Context Limit131,072Max Out: 65,536 tokens
Prompt Caching50% OFF$0.04 / 1M cached
Standard API / 1M$0.38🧠 Deep Reasoner
Model Capabilities & Contract SLAs
- Reasoning
- 🧠 Extended Test-Time Reasoning
gpt-oss-safeguard-20b is a safety reasoning model from OpenAI built upon gpt-oss-20b. This open-weight, 21B-parameter Mixture-of-Experts (MoE) model offers lower latency for safety tasks like content classification, LLM filtering, and trust...
Granular Pricing Matrix
Input Tokens (Standard Prompt)$0.07 / 1M
⚡ Cached Input (Prompt Cache Read)50% OFF$0.04 / 1M
Output Tokens (Completion)$0.30 / 1M
Pricing data via OpenRouter. Sync: 10/2/2026
Evaluate Competitors
VS Engine Matchupgpt-oss-safeguard-20b vs Thinking Machines: Inkling Small (free)VS Engine Matchupgpt-oss-safeguard-20b vs NVIDIA: Nemotron 3.5 Lightning (free)VS Engine Matchupgpt-oss-safeguard-20b vs Poolside: Laguna S 2.1VS Engine Matchupgpt-oss-safeguard-20b vs Thinking Machines: Inkling (free)VS Engine Matchupgpt-oss-safeguard-20b vs ByteDance: UI-TARS 7B VS Engine Matchupgpt-oss-safeguard-20b vs Qwen: Qwen3.8 27B (free)